Estimations for the production cross sections of medical 61, 64, 67Cu radioisotopes by using bayesian regularized artificial neural networks in (p, α) reactions

Abstract

Copper (Cu), which is produced in cyclotrons or reactors, is a significant tracer in the human body. Bayesian regularized artificial neural networks (ANNs) algorithm, which is one of the ANN approaches, was used in analyzing the production cross sections of 61Cu, 64Cu, and 67Cu radioisotopes in (p,α) reactions in the present study. The production cross sections of 61Cu, 64Cu, and 67Cu radioisotopes in (p,α) reactions were assessed by making use of the ANN algorithm and TALYS 1.95 codes. The estimated cross section data were then compared to the data found in the TALYS-Based Evaluated Nuclear Reaction Library 2019 (TENDL) and Experimental Nuclear Reaction Data (EXFOR) Library. ANN results were shown to yield successful correlation coefficients of 0.99477, 0.98665, and 0.99313 for training, testing, and all processes, respectively. Furthermore, the mean square error (MSE) results of ANN prediction were calculated to be 3.6 (training) and 11.84 (testing) mb for all the (p,α.) reactions. It was concluded that the ANN algorithm yielded successful results since ANN estimations were suitable for experimental data, TALYS 1.95 calculations, and TENDL data.